Most undervalued
Ranked by the gap between our base-case fair value and the share price, widest first. The first five show the revenue path behind the valuation and the assumption that makes our view different from the market.
A model is an explicit set of assumptions, not a price target. Open any chart to inspect the drivers, change the case, and see what would make the gap close or disappear.
Investors worry Uber's growth will slow after 2028 as ride-hailing and delivery competition intensifies, while autonomous vehicles could weaken its position as the middleman between riders and drivers. The Delivery Hero acquisition also adds integration risk just as the core business is expected to mature. That caution is visible in the bear case, where bookings slow sharply and the shares fall below today's price. Our base case is more optimistic: it assumes Uber's scale keeps customers and drivers on the platform, Delivery Hero expands the addressable market, and margins improve even as growth becomes steadier. The valuation gap exists because the market appears to be pricing in a much faster slowdown than the model does.
Open the full model →The market is paying 5.6x this model's FY2027 EBITDA for a company that just printed an 84.6% gross margin and guided the next quarter up another 18%, and the reason is written into the guidance itself: Sandisk's volume is capped. It owns no fab, takes roughly half of Flash Ventures' output, and grows bits mid-to-high teens because that is how fast node transitions raise density. Everything else that happened in FY2026 was price - revenue per gigabyte up almost 150% in Datacenter, almost 180% in Edge, low-fifties percent in Consumer - and management's own FY2028-FY2030 framework asks the market to treat that shortage price as a plateau. The stock's behaviour says it does not: it peaked near $2,354 in late June 2026 and lost about 55% in a month. That is the whole disagreement, and it is one number - where revenue per gigabyte settles once the industry's own capital spending lands after calendar 2027.
Open the full model →Super Micro carries more than an ordinary hardware-cycle discount. Investors worry its AI-server boom could fade as competition increases, but they are also pricing in China and reporting risk. A 2026 export-control case alleged that servers were routed through Taiwan into China; Super Micro was not charged, but the company opened an independent review. That arrived after its former auditor resigned in 2024, financial filings were delayed, and the company faced a Nasdaq delisting deadline. The missing reports were later filed and an independent committee said it found no evidence of misconduct, yet the episode left a lasting trust discount. The bear case assumes rack growth slows and recent margins prove temporary. Our base case does not require better margins; it assumes AI infrastructure demand lasts long enough to fill more capacity. The gap reflects both fear of a post-boom decline and doubt that execution and reporting will remain clean.
Open the full model →Upstart is still priced like a leveraged bet on interest rates and consumer credit. When borrowing costs rise or defaults increase, lenders buy fewer loans and Upstart's marketplace can slow abruptly. Investors are also cautious because the company only recently returned to profit and still carries meaningful balance-sheet lending exposure. The bear case assumes the credit environment worsens just as loan growth cools. Our base case assumes the opposite is not necessary: credit can remain merely stable while auto and home lending grow from small bases and the core personal-loan business keeps recovering. The gap exists because the market is assigning a large macroeconomic discount to earnings that the model believes are becoming more diversified.
Open the full model →Hims & Hers carries a real regulatory discount. Its weight-loss growth relied partly on compounded GLP-1 treatments, which face tighter limits now that branded drugs are more available. The FDA previously warned the company over the sale and marketing of unapproved and allegedly misbranded compounded drugs, later named Hims while announcing further enforcement, and HHS referred potential violations to the Department of Justice. No criminal charges had been filed when this model was prepared, so this is an unresolved overhang rather than established liability. Investors must still price the possibility of fines, operating restrictions, or a faster shift toward lower-margin branded drugs. The bear case captures that fear: subscribers keep growing, but regulation and drug costs prevent profits from following. Our base case assumes weight-loss margins stabilise while established specialties and newer services carry more of the growth. The valuation gap reflects a disagreement about whether Hims can become a broader healthcare platform before regulatory pressure changes its economics.
Open the full model →